Maps are often used to convey information generated by models, for example, modeled cancer risk from air pollution. The concrete nature of images, such as maps, may convey more certainty than warranted for modeled information. Three map features were selected to communicate the uncertainty of modeled cancer risk: (a) map contours appeared in or out of focus, (b) one or three colors were used, and (c) a verbal-relative or numeric risk expression was used in the legend. Study aims were to assess how these features influenced risk beliefs and the ambiguity of risk beliefs at four assigned map locations that varied by risk level. We applied an integrated conceptual framework to conduct this full factorial experiment with 32 maps that varied by the three dichotomous features and four risk levels; 826 university students participated. Data was analyzed using structural equation modeling. Unfocused contours and the verbal-relative risk expression generated more ambiguity than their counterparts. Focused contours generated stronger risk beliefs for higher risk levels and weaker beliefs for lower risk levels. Number of colors had minimal influence. The magnitude of risk level, conveyed using incrementally darker shading, had a substantial dose-response influence on the strength of risk beliefs. Personal characteristics of prior beliefs and numeracy also had substantial influences. Bottom-up and top-down information processing suggest why iconic visual features of incremental shading and contour focus had the strongest visual influences on risk beliefs and ambiguity. Variations in contour focus and risk expression show promise for fostering appropriate levels of ambiguity.

A pilot study was conducted using an occupied, single-family test house in Columbus, OH, to determine whether a script-based protocol could be used to obtain data useful in identifying the key factors affecting air-exchange rate (AER) and the relationship between indoor and outdoor concentrations of selected traffic-related air pollutants. The test script called for hourly changes to elements of the test house considered likely to influence air flow and AER, including the position (open or closed) of each window and door and the operation (on/off) of the furnace, air conditioner, and ceiling fans. The script was implemented over a 3-day period (January 30-February 1, 2002) during which technicians collected hourly-average data for AER, indoor, and outdoor air concentrations for six pollutants (benzene, formaldehyde (HCHO), polycyclic aromatic hydrocarbons (PAH), carbon monoxide (CO), nitric oxide (NO), and nitrogen oxides (NO x )), and selected meteorological variables. Consistent with expectations, AER tended to increase with the number of open exterior windows and doors. The 39 AER values measured during the study when all exterior doors and windows were closed varied from 0.36 to 2.29 h À1 with a geometric mean (GM) of 0.77 h À1 and a geometric standard deviation (GSD) of 1.435. The 27 AER values measured when at least one exterior door or window was opened varied from 0.50 to 15.8 h À1 with a GM of 1.98 h À1 and a GSD of 1.902. AER was also affected by temperature and wind speed, most noticeably when exterior windows and doors were closed. Results of a series of stepwise linear regression analyses suggest that (1) outdoor pollutant concentration and (2) indoor pollutant concentration during the preceding hour were the ''variables of choice'' for predicting indoor pollutant concentration in the test house under the conditions of this study. Depending on the pollutant and ventilation conditions, one or more of the following variables produced a small, but significant increase in the explained variance (R 2 -value) of the regression equations: AER, number and location of apertures, wind speed, air-conditioning operation, indoor temperature, outdoor temperature, and relative humidity. The indoor concentrations of CO, PAH, NO, and NO x were highly correlated with the corresponding outdoor concentrations. The indoor benzene concentrations showed only moderate correlation with outdoor benzene levels, possibly due to a weak indoor source. Indoor formaldehyde concentrations always exceeded outdoor levels, and the correlation between indoor and outdoor concentrations was not statistically significant, indicating the presence of a strong indoor source.

No abstract

Polyurethane products were subjected to chamber testing to determine their emission rates of 2,4- and 2,6-toluene diisocyanate (TDI). The polyurethane (PU) products included carpet padding, furniture cushions, sheet foam, varnishes, and sealants, as well as a commercially-applied water sealant product for concrete that contained up to 4 percent TDI by weight. The PU products were screened in a 9-L glass chamber, under elevated temperature and chamber loading conditions, using both a time-integrated sampling and analysis method specific for TDI and a continuous but non-specific real-time monitor for isocyanates. None of the products normally found in residences showed a positive response in the screening tests, indicating that TDI emissions and consequently toxic effects from such products are negligible. However, the commercially-applied water sealant gave a positive response in the screening test. Further testing of that product at realistic temperatures showed initial TDI emission rates of about 300,000 micrograms/m2/hr, with emissions lasting only one hour or less. At 21 and 27 degrees C, about 1 percent and 5 percent, respectively, of the TDI content of the product was released to the air. The emitted TDI was predominantly the 2,6-isomer, although the TDI originally present in the product was predominantly the 2,4-isomer.

No abstract